Title: 2 Vaccine Produces Protective Adaptive Immunity In ... · 03/09/2020 · Immune gene expression following LUNAR-COV19 and conventional mRNA vaccination . 123 . C57BL/6J mice

1

Title: A Single Dose of Self-Transcribing and Replicating RNA Based SARS-CoV-2 1 Vaccine Produces Protective Adaptive Immunity In Mice. 2 3

Authors: 4

Ruklanthi de Alwis1,2*, Esther S Gan2*, Shiwei Chen2, Yan Shan Leong1, Hwee Cheng Tan2, 5

Summer L Zhang2, Clement Yau2, Daiki Matsuda3, Elizabeth Allen3, Paula Hartman3, Jenny 6

Park3, Maher Alayyoubi3, Hari Bhaskaran, Adrian Dukanovic, Belle Bao3, Brenda 7

Clemente3, Jerel Vega3, Scott Roberts3, Jose A. Gonzalez3, Marciano Sablad3, Rodrigo 8

Yelin3, Wendy Taylor3, Kiyoshi Tachikawa3, Suezanne Parker3, Priya Karmali3, Jared 9

Davis3, Sean M. Sullivan3, Steve G. Hughes3, Pad Chivukula3, Eng Eong Ooi1,2 10

11

12 1Viral Research and Experimental Medicine Center, SingHealth Duke-NUS Academic 13

Medical Center, Singapore. 14 2Program in Emerging Infectious Diseases, Duke-NUS Medical School, Singapore. 15 3Arcturus Therapeutics, Inc., 10628 Science Center Drive, San Diego CA 92121 16

17

*Equal contribution 18

Correspondence to: Dr. Eng Eong Ooi; [email protected] 19

20

21

22

23

Key words 24

SARS-CoV-2, conventional mRNA, self-amplifying RNA, STARR®, LUNAR®-COV19, 25

COVID-19, Vaccine, Coronavirus 26

27

28

29

30

31

32 33

34

preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted September 3, 2020. ; https://doi.org/10.1101/2020.09.03.280446doi: bioRxiv preprint

https://doi.org/10.1101/2020.09.03.280446

2

ABSTRACT 35

A self-transcribing and replicating RNA (STARRTM) based vaccine (LUNAR®-COV19) has 36

been developed to prevent SARS-CoV-2 infection. The vaccine encodes an alphavirus-based 37

replicon and the SARS-CoV-2 full length spike glycoprotein. Translation of the replicon 38

produces a replicase complex that amplifies and prolong SARS-CoV-2 spike glycoprotein 39

expression. A single prime vaccination in mice led to robust antibody responses, with 40

neutralizing antibody titers increasing up to day 60. Activation of cell mediated immunity 41

produced a strong viral antigen specific CD8+ T lymphocyte response. Assaying for 42

intracellular cytokine staining for IFN-γ and IL-4 positive CD4+ T helper lymphocytes as 43

well as anti-spike glycoprotein IgG2a/IgG1 ratios supported a strong Th1 dominant immune 44

response. Finally, single LUNAR-COV19 vaccination at both 2 µg and 10 µg doses 45

completely protected human ACE2 transgenic mice from both mortality and even measurable 46

infection following wild-type SARS-CoV-2 challenge. Our findings collectively suggest the 47

potential of Lunar-COV19 as a single dose vaccine. 48

49

50


https://doi.org/10.1101/2020.09.03.280446

3

INTRODUCTION 51

The pandemic of coronavirus disease-2019 (COVID-19) has afflicted tens of millions of 52

people, of which hundreds of thousands have died from severe respiratory dysfunction and 53

other complications of this disease [1]. The etiological agent of COVID-19 is the severe acute 54

respiratory syndrome coronavirus 2 (SARS-CoV-2), which may have first emerged from a 55

zoonotic source to then spread from person-to-person until global dissemination [1]. Current 56

control measures to curb the pandemic, such as national lockdowns, closure of work places 57

and schools and reduction of international travel are threatening to draw the world into a 58

global economic recession of unprecedented scale [2]. Vaccines that elicit durable protection 59

against SARS-CoV-2 infection are thus urgently needed [3]. Encouragingly, hundreds of 60

different vaccine development efforts are currently in progress, some of which have even 61

entered phase III clinical trials [4, 5]. 62

63

Despite some candidates reaching late-stage clinical trials, there is some uncertainty that 64

production can be upscaled in a sufficiently accelerated timeline to manufacture the billions 65

of vaccine doses required to immunize the world’s population [6]. Furthermore, recent results 66

from early phase COVID-19 vaccine trials have suggested that more than one dose would be 67

needed to elicit reasonable levels of adaptive immune memory [7-9]. Durable protection with 68

a single dose has been achieved with some viral live-attenuated vaccines (LAV), such as the 69

yellow fever vaccine [10-12]. However, since the genetic determinants of the clinical fitness 70

of SARS-CoV-2 are not well defined, development of a LAV SARS-CoV-2 strain that is safe 71

for use in humans is challenging. An alternative approach would be to mimic the key 72

immunogenic properties of live viral vaccines, to develop an alternate vaccine platform that 73

could also be effective in preventing COVID-19 with a single dose. A single dose vaccine 74

would not only avoid logistics and compliance challenges associated with multi-dose 75

vaccines, but also allow vaccination of more individuals with each batch [6]. 76

77

RNA vaccines offer a rapid approach to develop a COVID-19 vaccine [13]. RNA vaccines 78

are designed using the genetic sequence of the viral antigen and rapidly manufactured using 79

cell-free, rapidly scalable techniques [14]. The RNA is encapsulated in a lipid nanoparticle 80

(LNP), which generates robust immune responses without the need for adjuvants [15, 16]. 81

There are two main categories of RNA vaccines; 1) the conventional messenger RNA 82

(conventional mRNA) vaccine, where the immunogen of interest is directly translated from 83

the input vaccine transcript, and 2) the newer self-replicating RNA (replicon) vaccines [14]. 84


https://doi.org/10.1101/2020.09.03.280446

4

Replicon vaccines encode replication machinery, usually alphavirus-based replication 85

complex, that amplify sub-genomic RNA carrying the antigen of interest, resulting in the 86

amplification of transcripts bearing the antigen by several orders of magnitude over the initial 87

dose [17]. Prolonged antigen expression by such a construct could not only produce the 88

obvious dose sparing effects [17] but potentially also elicit innate and adaptive immune 89

responses similar to those associated with live vaccines. Herein, we show a head-to-head 90

comparison between a self-replicating RNA vaccine using Arcturus’ proprietary Self-91

Transcribing and Replicating RNA (STARRTM technology and a conventional mRNA 92

vaccine against SARS-CoV-2 and suggest that the STARR vaccine, LUNAR-COV19 offers 93

superior vaccine-induced immune responses to conventional mRNA. 94

95

RESULTS 96

97

Comparison of design and expression of STARR and conventional mRNA platforms 98

Both LUNAR-COV19 and conventional mRNA vaccine constructs were designed to encode 99

the full-length, unmodified, pre-fusion SARS-CoV-2 S protein (1273 aa), with LUNAR-100

COV19 additionally encoding the Venezuelan equine encephalitis virus (VEEV) replicase 101

genes required for self-amplification (Figure 1A). We first defined the characteristics of 102

these different constructs, which were both formulated with the same LUNAR LNP lipid 103

formulation. Despite differences in RNA lengths for LUNAR-COV19 and conventional 104

mRNA , the LNP diameter, polydispersity index and RNA trapping efficiency were similar 105

(Figure 1B). In vitro expression of the LUNAR-COV19 and conventional mRNA vaccine 106

were confirmed in cell lysate 24 hours post-transfection through positive western blot 107

detection of the S protein (Figure 1C). It was also observed that both vaccines expressed a 108

mixture of full-length S protein and cleaved S protein, i.e. into S1 and S2 transmembrane and 109

cytoplasmic membrane domains (Figure 1C). We then compared in vivo protein expression 110

of the two RNA platforms in BALB/c mice, by using STARR and conventional mRNA 111

constructs that expressed a luciferase reporter (Figure 1D). As expected, animals injected 112

with the conventional mRNA vaccine construct showed high in vivo luciferase expression at 113

day 1 although the expression levels declined significantly three days post injection. In 114

contrast, the luciferase expression in STARR injected mice showed increased signal of 115

protein production compared to conventional mRNA at all time points after Day 1 up to Day 116

7 post-inoculation (the last time point measured) and at doses ≥2.0 µg, protein expression 117


https://doi.org/10.1101/2020.09.03.280446

5

appeared to be still rising at day 7 (Figure 1D). These data showed that dose-for-dose, the 118

STARR luciferase construct yielded higher and more prolonged duration of luciferase 119

expression compared to mice injected with the conventional mRNA luciferase construct. 120

121

Immune gene expression following LUNAR-COV19 and conventional mRNA vaccination 122

C57BL/6J mice were vaccinated with LUNAR-COV19 or conventional mRNA vaccines at 123

0.2 µg, 2 µg and 10 µg doses or PBS control. No significant mean loss in animal weight 124

occurred over the first 4 days, except for those that received 10 µg of LUNAR-COV19 125

(Figure 2A). However, apart from weight loss, there were few other clinical signs as 126

indicated by the minimal differences in clinical scores. Both weight and clinical scores 127

improved uneventfully after day 3 post vaccination. 128

129

The innate immune response, particularly the type-I interferon (IFN) response has previously 130

been shown to be associated with vaccine immunogenicity following yellow fever 131

vaccination [11, 12, 18]. Furthermore, we have also found that reactive oxygen species-132

driven pro-inflammatory responses underpinned systemic adverse events in yellow fever 133

vaccination [19, 20]. Therefore, we measured the expression of innate immune and pro-134

inflammatory genes in whole blood of C57BL/6 mice inoculated with either PBS, 135

conventional mRNA vaccine or LUNAR-COV19. Genes in the type-I IFN pathway were the 136

most highly expressed in animals inoculated with LUNAR-COV19 compared to either 137

conventional mRNA vaccine or PBS (Figure 2B and Supplementary Figure 1). By contrast, 138

genes associated with pro-inflammatory responses were mostly reduced in abundance 139

following LUNAR-COV19 vaccination compared with either conventional mRNA vaccine or 140

PBS (Figure 2B and Supplementary Figure 1). 141

142

Since adaptive immune responses develop in germinal centers in the draining lymph nodes, 143

we dissected the draining lymph nodes at day 7 post-inoculation (study schematic in Figure 144

2A). The inguinal lymph nodes of mice inoculated with LUNAR-COV19 showed a dose-145

dependent increase in weight, unlike those from mice inoculated with either conventional 146

mRNA vaccine or PBS; the mean weight of lymph nodes from mice given 10 µg of LUNAR-147

COV19 was significantly higher than those given the equivalent conventional mRNA vaccine 148

(Figure 2C). Principal component analysis (PCA) of immune gene expression showed 149

clustering of responses to each of the 3 doses of LUNAR-COV19 away from the PBS control 150


https://doi.org/10.1101/2020.09.03.280446

6

(depicted as red and orange spheres in Figure 2D-F), indicating clear differences in immune 151

gene expression between LUNAR-COV19 vaccinated and placebo groups. These trends were 152

also dissimilar to those from mice given conventional mRNA vaccine where at all tested 153

doses, the PCA displayed substantial overlap with PBS control (shown as blue and orange 154

spheres in Figure 2D-F). 155

156

We next assessed the differentially expressed genes in the lymph nodes of mice given 157

LUNAR-COV19 compared to those inoculated with mRNA vaccine. Volcano plot analysis 158

identified significant upregulation of several innate, B and T cells genes in LUNAR-COV19 159

immunized animals (Figure 2G-I). Some of the most highly differentially expressed genes 160

included, GZMB (required for target cell killing by cytotoxic immune cells) [21], S100A8 161

and S100A9 (factors that regulate immune responses via TLR4) [22], TNFRSF17 (also 162

known as BCMA and regulates humoral immunity) [23], CXCR3 (chemokine receptor 163

involved in T cell trafficking and function) [24] and AICDA (mediates antibody class 164

switching and somatic hypermutation in B cells) [25]. These findings collectively indicate 165

that the adaptive immune responses in the draining lymph nodes of mice inoculated with 166

LUNAR-COV19 may differ to those given the non-replicating mRNA vaccine. 167

168

LUNAR-COV19 induced robust T cell responses 169

We next investigated the cellular immune response following vaccination of C57BL/6 mice 170

(n=5 per group) with LUNAR-COV19 or conventional mRNA. At day 7 post-vaccination, 171

spleens were harvested and assessed for CD8 and CD4 T cells by flow-cytometry. The CD8+ 172

T cell CD44+CD62L- effector/memory subset was significantly expanded in LUNAR-173

COV19 vaccinated mice compared to those given either PBS or conventional mRNA vaccine 174

(Figure 3A-B). There was no statistically significant difference in the proportion of CD4+ T 175

effector cells in these animals (Figure 3C). IFNγ+ CD8+ T cells (with 2 µg and 10 µg doses) 176

and IFNγ+ CD4+ T cells (in 0.2 µg and 10 µg) were proportionately higher, as found using 177

intracellular staining (ICS) with flow cytometry, in LUNAR-COV19 as compared to 178

conventional mRNA vaccinated animals (Figure 3D-F). 179

180

SARS-CoV-2 specific cellular responses were assessed in vaccinated animals by ELISPOT. 181

A set of 15-mer peptides covering the full length SARS-CoV-2 S protein were divided into 4 182

pools and tested for IFNγ+ responses in splenocytes of vaccinated and non-vaccinated 183


https://doi.org/10.1101/2020.09.03.280446

7

animals. SARS-CoV-2-specific cellular responses (displayed as IFNγ+ SFU/106 cells) were 184

detected by ELISPOT in both LUNAR-COV19 and conventional mRNA vaccine immunized 185

animals compared to PBS control (Figure 3G-I). These responses were substantially higher 186

across all doses in LUNAR-COV19 compared to conventional mRNA vaccinated groups 187

(Figure 3G-I). Even the highest tested dose (10 µg) of conventional mRNA vaccine 188

produced IFNγ+ ELISPOT responses that were appreciably lower than those by the lowest 189

dose (0.2 µg) of LUNAR-COV19. 190

191

LUNAR-COV19 induced superior humoral immune responses 192

SARS-CoV-2-specific humoral responses following vaccination with a single injection were 193

characterized in two different mouse models, BALB/c and C57BL/6. Female mice (n=5 per 194

group) were vaccinated at day 0 and bled every 10 days, up to day 60 for BALB/c and day 30 195

for C57BL/6 (Figure 4A). SARS-CoV-2 S-specific IgM responses were tested at 1:2000 196

serum dilution using an in-house Luminex immuno-assay. All tested doses of the 197

conventional mRNA vaccine and LUNAR-COV19 produced detectable S-specific IgM 198

responses in both mouse models (Figure 4B-C). When comparing LUNAR-COV19 to 199

conventional mRNA vaccinated BALB/c mice, no difference in IgM responses was observed; 200

IgM levels in C57BL/6 mice were higher in LUNAR-COV19 vaccinated C57BL/6 mice at 201

day 10 post vaccination. In contrast, SARS-CoV-2 S-specific IgG (at 1:2000 serum dilution) 202

levels were universally higher from day 20 onwards in animals inoculated with LUNAR-203

COV19 compared to conventional mRNA vaccine (Figure 4D-E). Perhaps even more 204

remarkably, in BALB/c vaccinated with LUNAR-COV19, the IgG levels continued to 205

increase until day 50 post-vaccination; C57BL/6 mice were only monitored until day 30 post-206

vaccination. This trend contrasted sharply with mice that received the conventional mRNA 207

vaccine where in BALB/c mice antibody levels plateaued after day 10 post-vaccination; 208

although increasing S-specific IgG levels were observed in conventional mRNA-vaccinated 209

C57BL/6 mice these were universally lower than those that received LUNAR-COV19. 210

211

In depth characterization of the SARS-CoV-2 specific IgG response in vaccinated animals 212

was conducted at day 30 post-immunization to assess which regions of S protein are targeted. 213

IgG endpoint titers were estimated to full ectodomain S protein, S1, S2 and receptor binding 214

domain (RBD) regions. As expected for both vaccine candidates the majority of SARS-CoV-215

2 specific IgG recognized S1, which contains the RBC, although high IgG endpoint titers 216


https://doi.org/10.1101/2020.09.03.280446

8

were also detected to S2 protein (Figure 4F-G). However, LUNAR-COV19 elicited IgG 217

endpoint titers were universally and significantly higher compared to those produced by 218

conventional mRNA vaccination (Figure 4F-G). Notably, IgG that bind the RBD of S 219

protein, which is an immunodominant site of neutralizing antibodies [26, 27], were also 220

higher in LUNAR-COV19 compared to conventional mRNA vaccinated animals. It is also 221

noteworthy that at lower doses, conventional mRNA vaccine but not LUNAR-COV19 222

struggled to elicit high SARS-CoV-2 specific IgG titers in the more Th1 dominant C57BL/6 223

mouse strain (Figure 4G). Taken collectively, a single dose of LUNAR-COV19 induced 224

significant differences in immune gene expression and superior cellular immune responses in 225

draining lymph nodes compared to the conventional mRNA vaccine and consequently greater 226

and more prolonged humoral immune responses. 227

228

We assessed both the binding strength (avidity) and the neutralizing ability of the antibody 229

response elicited by these vaccine constructs. Serum IgG avidity was measured at day 30 230

post-vaccination using a modified Luminex immuno-assay with 8M urea washes. LUNAR-231

COV19 elicited higher avidity S protein-specific IgG in both mouse models at all tested 232

doses (Figure 4H). These differences were observed, with the exception of 0.2 µg in 233

BALB/c, across all doses (Figure 4H), indicating that LUNAR-COV19 elicited better quality 234

antibodies, suggesting superior affinity maturation with the LUNAR-COV19 vaccine. 235

236

Neutralization of live SARS-CoV-2 by serum from vaccinated animals was assessed using 237

the plaque reduction neutralization test (PRNT). At day 30 LUNAR-COV19 vaccinated 238

BALB/c mice showed a clear dose-dependent elevation in PRNT50 titers; 4 out of 5 (80%) of 239

mice in the 10 µg LUNAR-COV19 group showed PRNT50 titers greater than 320, which was 240

the upper limit of our dilution (Figure 4I). Similar dose-dependent trends in PRNT50 titers 241

were also found in C57BL/6 mice although in these animals, the PRNT50 titers of several 242

animals exceeded 320 even with the lowest 0.2 µg dose vaccination (Figure 4I). In sharp 243

contrast, PRNT50 titers in animals inoculated with the conventional mRNA vaccine construct 244

were, except for one C57BL/6J mouse that received 10 µg dose, all <20 (Figure 4I). 245

Unexpectedly but encouragingly, PRNT50 and PRNT70 titers of LUNAR-COV19 vaccinated 246

BALB/c mice continued to rise between day 30 and day 60 after a single vaccination (Figure 247

4J-K) and at both time points for doses ≥2.0 µg. These titers were comparable to PRNT70 248

titers for sera from convalescent COVID-19 patients (Figure 4K). 249


https://doi.org/10.1101/2020.09.03.280446

9

250

We also found that the S protein IgG titers positively correlated with PRNT50 titers with 251

LUNAR-COV19 vaccinated mice in both mouse models (Figure 4L). Similar positive 252

correlations were also observed with IgG against S1 and RBD (Supplementary Figure 1). 253

By contrast, we found no correlation between IgG and PRNT50 titers in conventional mRNA 254

vaccinated mice (Figure 4L). Taken collectively, our antibody response analyses suggest that 255

the higher PRNT50 titers following vaccination with LUNAR-COV19 are not only strongly 256

associated with the amount of IgG produced but are also a factor of the superior quality of the 257

anti-SARS-CoV-2 antibodies produced following vaccination with LUNAR-COV19. 258

259

LUNAR-COV19 vaccination showed a Th1 dominant response 260

A safety concern for a coronavirus vaccine is the risk of vaccine-associated immune 261

enhancement of respiratory disease (VAERD) [28]. Indeed, SARS-CoV and MERS-CoV 262

vaccine development have highlighted the importance of Th1 skewed responses in mitigating 263

the risk of vaccine-induced immune enhancement [29, 30]. Therefore, we investigated the 264

Th1/ Th2 balance elicited by vaccination with both conventional mRNA and LUNAR-265

COV19. The IgG subclass fate of plasma cells are highly influenced by T helper (Th) cells 266

[31]. At day 30 post-vaccination, both conventional mRNA and LUNAR-COV19, induced 267

comparable amounts of SARS-CoV-2 S-specific IgG1, a Th2-associated IgG subclass in 268

mice, except for the 0.2 µg dose in C56BL/6J mice (Figure 5A-B). In contrast, the Th1-269

associated IgG subclasses - IgG2a in BALB/c and IgG2c in C56BL/6J - were significantly 270

greater in LUNAR-COV19 vaccinated animals. The ratios of S protein-specific IgG2a/IgG1 271

(Balb/c) and IgG2c/IgG1 (C57BL/6) were greater than 1 in LUNAR-COV19 vaccinated 272

animals (Figure 5A-B). Except for the 0.2 µg dose, these ratios were all significantly greater 273

with LUNAR-COV19 compared to the conventional mRNA vaccinated animals. 274

275

Additionally, we used ICS to investigate the production of IFNγ (Th1 cytokine) and IL4 (Th2 276

cytokine) by CD4+ T cells in spleens at day 7 post vaccination C56BL/6J mice. As was 277

described above, compared to conventional mRNA vaccination, IFNγ levels were 278

significantly greater in LUNAR-COV19 vaccinated animals (Figure 3F). IL4 expression in 279

CD4 T cells were slightly higher with conventional mRNA as compared to LUNAR-COV19 280

at 0.2 and 2.0 µg doses (Figure 5C). In comparing the IFNγ and IL4 levels in individual 281

mice, we found that the ratios of IFNγ/IL4 in CD4+ T cells for both LUNAR-COV19 and 282


https://doi.org/10.1101/2020.09.03.280446

10

conventional mRNA vaccinated mice were universally above 1 (Figure 5D). The ratio of 283

IFNγ/IL4 in CD4+ T cells in mice given the 0.2 and 2.0 µg doses were significantly greater 284

with LUNAR-COV19 than conventional mRNA vaccination (Figure 5F). However, the 285

elevated ratios at these doses were due to a decrease in IL4 expression at levels below 286

background (i.e. PBS control mice), rather than reduced IFNγ and hence Th1 activity. Taken 287

collectively, our data show that LUNAR-COV19 produced a Th1 biased adaptive immune 288

response. 289

290

Single dose of LUNAR-COV19 protects from a lethal infection of SARS-CoV-2 291

Finally, we tested the efficacy of LUNAR-COV19 in protecting against infection and 292

mortality in a lethal SARS-CoV-2 challenge model. Transgenic hACE2 mice immunized 293

with either PBS, or 2 µg or 10 µg of LUNAR-COV19 vaccine were intranasally challenged 294

with live SARS-CoV-2 virus (5x104TCID50) at day 30 post-vaccination. This was the same 295

isolate as that used for our PRNT assays. Mice were then divided into two groups: one group 296

was tracked for weight, clinical scores and survival; a second group of mice were euthanized 297

at 5 days post injection (dpi) and viral loads assessed in the respiratory tract (trachea to lung) 298

and brain (Figure 6A). Measurement of PRNT70 titers confirmed the generation of 299

neutralizing antibodies in LUNAR-COV19-vaccinated hACE2 mice (Figure 6B). 300

Irrespective of tested dosages, mice that received the LUNAR-COV19 vaccine showed 301

unchanged weight and no clinical sign, while the PBS mice showed significant drop in 302

weight and increased clinical scores upon challenge with wild-type SARS-CoV-2 (Figure 303

6C-D). LUNAR-COV19 vaccination at both 2 µg and 10 µg doses fully protected hACE2 304

mice from an otherwise 100% mortality at day 7 post-challenge (Figure 6E). Assessment of 305

tissue viral load at day 5 post-challenge found minimal to no SARS-CoV-2 RNA (Figure 6F) 306

in contrast to unvaccinated animal controls. Although viral RNA was detectable at very low 307

levels in some animals, this was not associated with any presence of infectious viral particles, 308

so most like represents viral RNA fragments rather than intact viral RNA genomes. No 309

detectable infectious virus was found in either the respiratory tracts or brains of LUNAR-310

COV19 vaccinated animals (Figure 6G). By contrast, unvaccinated animals showed 4 and 8 311

logs of infectious SARS-CoV-2 in the respiratory tract and brain, respectively (Figure 6G). 312

Collectively, these data show that a single dose of LUNAR-COV19 vaccine induced robust 313

humoral and cellular immune responses that led to complete protection of hACE2 mice from 314

a lethal SARS-CoV-2 challenge. 315


https://doi.org/10.1101/2020.09.03.280446

11

316

DISCUSSION 317

The pandemic of COVID-19 has necessitated rapid development of vaccines. Encouragingly, 318

several COVID-19 vaccine candidates are now in clinical trials and more are entering first-in-319

human trials. However, the majority of vaccine candidates being developed require two or 320

more doses for sufficient adaptive immune responses. Requirement for a second shot could 321

complicate compliance rate in mass vaccination campaigns and results in fewer subjects 322

vaccinated per batch, thereby reducing the efficiency of vaccination. Hence, a single dose 323

vaccine that generates robust and sustained cellular and humoral immunity, without elevating 324

the risk of vaccine-mediated immune enhancement, remains an unmet need. 325

326

Amongst the licensed vaccines for other diseases, live attenuated vaccines can offer the most 327

durable protection against viral diseases. Live vaccines infect and replicate at sites of 328

inoculation and some even in draining lymph nodes. Replication enables endogenous and 329

sustained expression of viral antigens that enable antigen presentation to stimulate cytotoxic 330

CD8+ T cells. Expressed antigens taken up by antigen presenting cells also trigger CD4+ T 331

cell help that drives affinity maturation in B cells. Studies on the live attenuated yellow fever 332

vaccine, have shown that a longer period of stimulation of the adaptive immune response 333

results in superior adaptive immune responses [32]. Although work to determine which of 334

these correlates of live vaccines are mechanistic determinants of adaptive immunity is still 335

ongoing, the ability of self-replicating RNA vaccines to simulate the sustained antigen 336

presentation characteristics of live vaccination could offer durable immunity against COVID-337

19. 338

339

Numerous studies have shown RNA vaccines to be immunogenic. In this study, we 340

conducted a side-by-side comparison of the immunogenicity elicited by two SARS-CoV-2 341

RNA vaccine candidates, a conventional mRNA construct and the STARR construct, 342

LUNAR-COV19. We found that, compared to conventional mRNA, LUNAR-COV19 343

produced higher and longer protein expression in vivo, upregulated the gene expression of 344

several innate, B and T cell response genes in the blood and draining lymph nodes. These 345

properties were associated with significantly greater neutralizing antibody and SARS-CoV-2 346

specific IgG responses, CD8+ T cell responses, IFNγ+ ELISPOT responses, and Th1 skewed 347

responses (which have been shown to associate with protection from VAERD) than 348


https://doi.org/10.1101/2020.09.03.280446

12

conventional mRNA. Interestingly, despite the highest tested dose of conventional mRNA 349

eliciting comparable S protein-specific antibodies as the lowest tested dose of LUNAR-350

COV19, the conventional mRNA-elicited IgG did not show such robust avidity or 351

neutralization activity as those from LUNAR-COV19 vaccination. These data suggest a 352

qualitatively better humoral immune response with superior affinity maturation of B-cells 353

with the LUNAR-COV19 vaccine. Our findings thus highlight the immunological advantages 354

of self-replicating RNA over conventional mRNA platforms. 355

356

The superior quality of immune responses elicited by LUNAR-COV19 over the conventional 357

mRNA vaccine construct could be attributable to multiple factors. Higher and longer 358

expression of immunogens produce better immunity [32], likely through better engagement 359

of T follicular helper cells and thereby leading to more diverse antibody targets and more 360

robust neutralizing antibody responses [33, 34]. Replication of LUNAR-COV19 results in the 361

formation of a negative-strand template for production of more positive-strand mRNA and 362

sub-genomic mRNA expressing the S transgene. Interaction between the negative- and 363

positive-strands forms a double stranded RNA (dsRNA) intermediate, which would interact 364

with TLR3 and RIG-I-like receptors to stimulate type 1 interferon responses [35-37], which 365

we and others have previously shown to correlate with superior adaptive immune responses 366

[11, 12, 18]. Production of IFNγ can also stimulate development of cytotoxic CD8+ T cells 367

[36]. Importantly, the S protein does contain human CD8+ T cell epitopes. As suggested by 368

recent findings on T cell responses to SARS-CoV-2 and other coronavirus infections [38-40], 369

the development of T cell memory could be important for long-term immunity. 370

371

It is unclear whether the VEEV nsP1-4 forming the replication complex contains any 372

immunogenic properties although mutations in the nsP proteins have been shown to affect the 373

induction of type I IFN [41]. Although unexplored in our current study, VEEV replicons have 374

also been shown to adjuvant immune responses at mucosal sites [42], further justifying the 375

use of STARR platform to develop a COVID-19 vaccine. 376

377

In conclusion, STARR vaccine platform as exemplified by LUNAR-COV19, offers an 378

approach to simulate key immunogenic properties of live virus vaccination and offers the 379

potential for an effective single-shot vaccination against COVID-19. 380

381


https://doi.org/10.1101/2020.09.03.280446

13

METHODS (Supplement 1) 382

Vaccine plasmid constructs and design 383

A human codon-optimized spike (S) glycoprotein gene of SARS-CoV-2 (GenBank 384

accession: YP_009724390) was cloned into plasmids pARM2922 and pARM2379 for 385

generation of SARS-CoV-2 Spike expressing STARR and conventional mRNA, respectively. 386

The STARR plasmid also encoded for the Venezuela equine encephalitis virus (VEEV) non-387

structural proteins nsP1, nsP2, nsP3 and nsp4, which together form the replicase complex that 388

bind to the sub-genomic promoter placed right before the S protein sequence. The cloned 389

portions of all plasmid constructs were verified by DNA sequencing. Plasmids were 390

linearized immediately after the poly(A) stretch and used as a template for in vitro 391

transcription reaction with T7 RNA polymerase. For LUNAR-CoV19 vaccine, the reaction 392

for RNA was performed as previously described [43] with proprietary modifications to allow 393

highly efficient co-transcriptional incorporation of a proprietary Cap1 analogue and to 394

achieve high quality RNA molecule of over 11,000-nt long the STARR mRNA. RNA was 395

then purified through silica column (Macherey Nagel) and quantified by UV absorbance. For 396

the conventional mRNA vaccine, the RNA was synthesized similarly but with 100% 397

substitution of UTP with N1-methyl-pseudoUTP. For both LUNAR-CoV19 and 398

conventional mRNA vaccines, the RNA quality and integrity were verified by 0.8-1.2% non-399

denaturing agarose gel electrophoresis as well as Fragment Analyzer (Advanced 400

Analytical). The purified RNAs were stored in RNase-free water at -80 °C until further use. 401

402

Vaccine lipid nanoparticles (LNPs) 403

LUNAR® nanoparticles encapsulating STARR™ were prepared by mixing an ethanolic 404

solution of lipids with an aqueous solution of STARR™ RNA. Lipid excipients (Arcturus 405

Therapeutics proprietary ionizable lipid, DSPC, Cholesterol and PEG2000-DMG) are 406

dissolved in ethanol at mole ratio of 50:10: 38.5:1.5 or 50:13:35.5:1.5. An aqueous solution 407

of the vaccine RNA is prepared in citrate buffer pH 4.0. The lipid mixture is then combined 408

with the vaccine RNA solution at a flow rate ratio of 1:3 (V/V) via a proprietary mixing 409

module. Nanoparticles thus formed are stabilized by dilution with phosphate buffer followed 410

by HEPES buffer, pH 8.0. Ultrafiltration and diafiltration (UF/DF) of the nanoparticle 411

formulation is then performed by tangential flow filtration (TFF) using modified PES hollow-412

fiber membranes (100 kDa MWCO) and HEPES pH 8.0 buffer. Post UF/DF, the formulation 413

is filtered through a 0.2 µm PES filter. An in-process RNA concentration analysis is then 414


https://doi.org/10.1101/2020.09.03.280446

14

performed. Concentration of the formulation is adjusted to the final target RNA concentration 415

followed by filtration through a 0.2 µm PES sterilizing-grade filter. Post sterile filtration, 416

bulk formulation is aseptically filled into glass vials, stoppered, capped, and frozen at -70 ± 417

10°C. Analytical characterization included measurement of particle size and polydispersity 418

using dynamic light scattering (ZEN3600, Malvern Instruments), pH, Osmolality, RNA 419

content and encapsulation efficiency by a fluorometric assay using Ribogreen RNA reagent, 420

RNA purity by capillary electrophoresis using fragment analyzer (Advanced Analytical), 421

lipid content using HPLC,. 422

In vitro transfection and immunoblot detection of spike protein 423

Hep3b cells (seeded in 6-well plates at a density of 7 X 105 cells/well, a day before) were 424

transfected with purified IVTs (2.5 μg conventional mRNA and 2.5 μg STARR ) by 425

Lipofectamine MessengerMax transfection reagent (Thermo Fisher Scientific) according to 426

the manufacturer’s instruction. The cells were harvested the next day with a hypotonic buffer 427

(10 mM Tris-HCl, 10 mM NaCl supplemented with protease inhibitor cocktail (Roche) ) 428

followed by sonication. Samples were deglycosylated followed by treatment with PNGase F 429

(New England Biolabs) according to the manufacture’s instruction. 430

The protein lysate (10 μg) was resolved on a 7.5% NuPAGE Tris-Acetate gel (Thermo Fisher 431

Scientific), and the spike protein expression was analyzed by LI-COR Quantitative Western 432

Blot system using a rabbit antibody detecting S1 (40150-T62-COV2, Sino Biologic) and a 433

mouse antibody for S2 region (GTX632604, GeneTex) along with appropriate secondary 434

antibodies (goat anti-rabbit 800 and goat anti-mouse 680). 435

Animal studies 436

BALB/c studies 437

All BALB/c mouse studies were approved by the Explora Biolabs IACUC and performed 438

under the Animal Care and Use Protocol number EB-17-004-003. A head-to-head 439

comparison of the protein expression of the conventional mRNA and STARR vaccine 440

platforms was conducted using conventional mRNA and STARR constructs expressing a 441

luciferase reporter gene. BALB/c mice (Jackson Laboratory) were intramuscularly (IM) in 442

the rectus femoris with conventional mRNA or STARR at doses of 0.2, 2 and 10 µg (n=3 443

mice/group). Expression of the conventional mRNA and STARR constructs were measured 444


https://doi.org/10.1101/2020.09.03.280446

15

at days 1, 3 and 7 post-inoculation through luciferase expression by imaging the mice for 445

bioluminescence. 446

447

Humoral responses to the SARS-CoV-2 Spike vaccine candidates were tested in Female 448

BALB/c mice (Jackson Laboratory) aged 8-10 weeks by IM immunization of the rectus 449

femoris with either conventional mRNA or LUNAR-COV19 at doses 0.2 µg, 2 µg, or 10 µg 450

(n=5 mice/group). Mice were bled at baseline and at 10, 19, 30, 40, 50- and 60-days post-451

vaccination to assess SARS-CoV-2 specific humoral immune responses. 452

453

C57BL/6 454

All C57BL/6 mouse studies were performed in accordance with protocols approved by the 455

Institutional Animal Care and Use Committee at Singapore Health Services, Singapore (ref 456

no.: 2020/SHS/1554). C57BL/6 mice purchased from inVivos were housed in a BSL-2 457

animal facility at Duke-NUS Medical School. Groups of 6-8 weeks old wild-type C57BL/6 458

female mice were vaccinated intramuscularly with either conventional mRNA or LUNAR-459

COV19 at doses 0.2 µg, 2 µg, or 10 µg. For transcriptomic and T cell studies, submandibular 460

bleeds were performed for whole blood at 24 hrs post-vaccination. Day 7 post-immunization, 461

mice were sacrificed at and inguinal lymph nodes and spleen harvested for investigation of 462

immune gene expression and T cell responses, respectively. Splenocyte suspensions for 463

measuring T cell responses were obtained by crushing spleen through a 70µm cell strainer 464

(Corning). Red blood cells were removed by lysis using BD PharmLyse reagent. For 465

antibody studies, another set of vaccinated 6-8 weeks old mice were bled at baseline and at 466

10, 20, and 30 days post-vaccination. 467

468

SARS-CoV-2 challenge experiments were conducted with female B6;SJL-Tg(K18-469

hACE2)2Prlmn/J mice purchased from Jackson laboratory. Groups of 6-8 weeks old wild-470

type C57BL/6 female mice were vaccinated intramuscularly with 100 µl LUNAR-COV19 at 471

doses of 2 µg, or 10 µg. Submandibular bleeds were performed for serum isolation to 472

determine antibody titers via PRNT 28 days post vaccination. Animal were infected with 473

5x104 TCID50 in 50µl via the intranasal route. Daily weight measurements and clinical 474

scores were obtained. Mice were sacrificed when exhibiting greater than 20% weight loss or 475

clinical score of 10. To assess organ viral loads, mice were sacrificed 5 days post infection 476

and harvested organs were frozen at -80oC. Whole lungs and brains were homogenized with 477


https://doi.org/10.1101/2020.09.03.280446

16

MP lysing matrix A and F according to manufacturer’s instructions in 1ml PBS. Homogenate 478

was used to assess both plaque titers and RNA extraction using TRIzol LS (Invitrogen). No 479

blinding was done for animal studies. 480

481

Gene expression of immune and inflammatory genes 482

Whole blood collected 1-day post-vaccination was lysed using BD PharmLyse reagent and 483

RNA extracted using Qiagen RNAeasy kit. Mouse lymph nodes collected from 7 days post 484

vaccination were homogenized and RNA extracted using trizol LS. RNA (50 ng) from whole 485

blood cells and lymph nodes were hybridized to the NanoString nCounter mouse 486

inflammation and immunology v2 panels (Nanostring Technologies), respectively. As 487

previously described [20, 44], RNA was hybridized with reconstituted CodeSet and 488

ProbeSet. Reactions were incubated for 24 hours at 65oC and ramped down to down to 4oC. 489

Hybridized samples were then immobilized onto a nCounter cartridge and imaged on a 490

nCounter SPRINT (NanoString Technologies). Data was analyzed using the nSolver 491

Analysis software (Nanostring Technologies) and Partek Genomics Suite. For normalization, 492

samples were excluded when percentage field of vision registration is <75, binding density 493

outside the range 0.1–1.8, positive control R2 value is <0.95 and 0.5 fM positive control is ≤2 494

s.d. above the mean of the negative controls. Background subtraction was performed by 495

subtracting estimated background from the geometric means of the raw counts of negative 496

control probes. Probe counts less than the background was floored to a value of 1. The 497

geometric mean of positive controls was used to compute positive control normalization 498

parameters. Samples with normalization factors outside 0.3–3.0 were excluded. The 499

geometric mean of housekeeping genes was used to compute the reference normalization 500

factor. Samples with reference factors outside the 0.10–10.0 range were also excluded. 501

Hierarchical clustering was performed with Partek Genomics Suite v6 on gene sets zScore 502

values by Euclidean dissimilarity and average linkage. 503

504

To identify DEGs between groups, Partek Genomics Suite Analysis v7 software was used to 505

analyse variance (ANOVA) with a cut off-of P < 0.05. Log2 Fold Changes generated were 506

used for volcano plots constructed using Prism v8.1.0 software. DEGs were identified by a 507

fold change cut-off of 2. Unsupervised principle component analysis was performed to 508

visualize variability between vaccinated and non-vaccinated animals with Partek genomics 509

suite analysis v7 software. PCA ellipsoids were drawn with a maximum density and 3 510

subdivisions. 511


https://doi.org/10.1101/2020.09.03.280446

17

512 Flow cytometry 513

Surface staining was performed on freshly-isolated splenocytes using the following panel of 514

antibodies and reagents: B220 (RA3-6B2), CD3 (17A2), CD4 (RM4-5), CD8α (53-6.7), 515

CD44 (IM7), CD62L (MEL-14) and DAPI. Intracellular cytokine staining was performed by 516

stimulating freshly-isolated splenocytes with 50 ng/ml PMA and 500 ng/ml ionomycin in the 517

presence of GolgiPlug (BD) for 6 h. After stimulation, surface staining of CD3, CD4 and 518

CD8a was performed followed by intracellular staining of IFN-γ (XMG1.2) and IL-4 519

(11B11). Data acquisition was performed on a BD LSRFortessa and analyzed using FlowJo. 520

521

ELISPOT 522

ELISPOT was performed using mouse IFN-γ ELISpotBASIC kit (Mabtech). A similar protocol 523

has been used for human SARS-CoV-2 samples [40]. In brief, 4 x 105 freshly-isolated 524

splenocytes were plated into PVDF-coated 96 well plates containing IFN-γ capture antibody 525

(AN18). Cells were stimulated with a 15-mer peptide library covering part of the S protein. 526

143 total peptides were divided into four pools and used at a final concentration of 1 µg/ml 527

per peptide. Negative control wells contained no peptide. Following overnight stimulation, 528

plates were washed and sequentially incubated with biotinylated IFN-γ detection antibody 529

(R4-6A2), streptavidin-ALP and finally BCIP/NBT. Plates were imaged using ImmunoSpot 530

analyzer and quantified using ImmunoSpot software. 531

532

Luminex Immuno-assay 533

Longitudinal assessment of binding antibody 534

Longitudinal IgM and IgG responses in BALB/c and C57BL/6 were measured using an in-535

house Luminex Immuno-assay. Similar Luminex Immuno-assays have been previously 536

described for antibody detection against SARS-CoV-2 antigens [45, 46]. Briefly, Magpix 537

Luminex beads were covalently conjugated to insect-derived HIS-tagged SARS-CoV-2 538

whole Spike protein (SinoBiologicals) using the ABC coupling kit (Thermo) as per 539

manufacturer’s instructions. Beads were then blocked with 1%BSA, followed by incubation 540

with serum (diluted at 1:2000 in block) for 1 hr at 37C. Beads are then washed and SARS-541

CoV-2 Spike-specific mouse antibodies were detected using the relevant biotinylated 542

secondary antibody (i.e. anti-mouse IgM-biotin and anti-mouse IgG-biotin (Southern 543

Biotech) for IgM and IgG assessment, respectively) with streptavidin-PE (Southern Biotech). 544

Antibody binding to Spike were then measured on a Magpix instrument as median 545


https://doi.org/10.1101/2020.09.03.280446

18

fluorescence intensity (MFI). Spike antigen quantity on beads were also probed with anti-546

6xHIS-PE antibodies and all MFI values were then corrected to Spike antigen quantity to 547

account for experiment to experiment variation. 548

549

IgG and IgG subclass endpoint titers 550

IgG endpoint titers to mammalian-derived SARS-CoV-2 Spike, Spike domain 1 (S1), spike 551

domain 2 (S2) and receptor binding domain (RBD) at day 30 sera post-immunization were 552

measured using Luminex ImmunoAssay. Assay was conducted as described above, with the 553

modification of serially diluting serum 10-fold from 200 to 2x108. Similarly, IgG subclass 554

endpoint titers (i.e. IgG1 and IgG2a in BALB/c and IgG1 and IgG2c in C57BL/6) were 555

measured against mammalian-derived SARS-CoV-2 Spike protein using serially diluted 556

mouse sera (5-fold from 200 to 3.1x106) and secondary antibodies anti-IgG1-biotin, anti-557

IgG2a-biotin or anti-IgG2b-biotin (Southern Biotech). Four parameter logistic (4PL) curves 558

were fitted to the measured MFI data from serially diluted sera, and three times the 559

background (i.e. 3x MFI with no serum) was used as a threshold cutoff to estimate endpoint 560

titers. 561

562

IgG Avidity 563

Avidity index of IgG to SARS-CoV-2 Spike protein at day 30 sera post-immunization was 564

estimated using the Luminex ImmunoAssay. Assay was conducted as described above with 565

the minor modification of following bead incubation with serum (diluted at 1:2000) with 566

either a 10 min PBS or 8M urea wash. Avidity Index was estimated by subtracting 567

background MFI from all sample values, and then dividing MFI with 8M Urea wash by MFI 568

with PBS wash. 569

570

Neutralization assay 571

Virus Neutralization titer assay (VNT) 572

Neutralization sero-conversion was assessed at day 10 and 20 post-immunization in BALB/c 573

using a virus neutralization assay as previously described [47]. Briefly, sera were diluted to 574

1:20 in culture media, mixed at a 1:1 ratio with a Singaporean clinical isolate of live SARS-575

CoV-2 virus, isolate BetaCoV/Singapore/2/2020 (GISAID accession code EPI_ISL_406973) 576

and incubated for 1 hr at 37C. Virus-antibody immune-complexes were then added to Vero-577

E6 cells (ATCC) in 96-well plates, and incubated at 37C. Five days later, plates were 578

assessed under a microscope for cytopathic effect (CPE) of the cells. 579


https://doi.org/10.1101/2020.09.03.280446

19

580

Plaque reduction neutralization titer (PRNT) 581

Neutralization of live SARS-CoV-2 was measured by PRNT at day 30 post-vaccination in 582

both BALB/c and C57BL/6 mice. Similar protocols have been published previously for 583

SARS-CoV-2 [48]. Briefly, mouse sera were serially diluted from 1:20 to 1:320 in culture 584

media and incubated with the Singapore isolate of SARS-CoV-2 virus for 1 hr at 37C. Virus-585

antibody mixtures were then added to Vero-E6 cells in 24-well plates, incubated for 1-2 hrs, 586

then overlayed with carboxymethyl cellulose (CMC) and incubated at 37C under 5% CO2. At 587

5 days, cells are washed, stained with crystal violet and plaques counted. The serum dilution 588

leading to neutralization of 50% of virus, i.e. PRNT50, was estimated. 589

590

ACKNOWLEDGMENTS 591

We thank the Economic Development Board of Singapore for initiating this collaboration and 592

for funding the development of LUNAR-COV19. ARdA received salary support from the 593

National Medical Research Council (NMRC) Young Investigator Award. EEO received 594

salary support from the NMRC Clinician-Scientist Award (Senior Investigator). 595

596

AUTHOR CONTRIBUTIONS 597

Ruklanthi de Alwis was responsible for the humoral characterization of the immune response 598

from the LUNAR COV19 vaccine. Esther S Gan was responsible for the huACE2 transgenic 599

mouse challenge studies and expression profiling analysis of the LUNAR COV19 vaccine. 600

DECLARATION OF INTERESTS 601

D.M., E.A., P.H., J.P., M.A., H.B., A.D., B.B., B.C., J.V., S.R, J.A.G., M.S., R.Y., W.T., 602

K.T., S.P., P.K., J.D., S.S., S.H. and P.C. are employees of Arcturus Therapeutics, Inc. 603

604

REFERENCES 605

1. WHO, WHO Coronavirus Disease (COVID-19) Dashboard. 2020. 606 2. Bank, W. The Global Economic Outlook During the COVID-19 Pandemic: A Changed 607

World. 2020; Available from: https://www.worldbank.org/en/news/feature/2020/06/08/the-608 global-economic-outlook-during-the-covid-19-pandemic-a-changed-world. 609

3. Randolph, H.E. and L.B. Barreiro, Herd Immunity: Understanding COVID-19. Immunity, 610 2020. 52(5): p. 737-741. 611

4. WHO. Draft landscape of COVID-19 candidate vaccines. 2020; Available from: 612 https://www.who.int/publications/m/item/draft-landscape-of-covid-19-candidate-vaccines. 613

5. Thanh Le, T., et al., The COVID-19 vaccine development landscape. Nat Rev Drug Discov, 614 2020. 19(5): p. 305-306. 615

6. Lurie, N., et al., Developing Covid-19 Vaccines at Pandemic Speed. N Engl J Med, 2020. 616 382(21): p. 1969-1973. 617


https://www.worldbank.org/en/news/feature/2020/06/08/the-global-economic-outlook-during-the-covid-19-pandemic-a-changed-world

https://www.worldbank.org/en/news/feature/2020/06/08/the-global-economic-outlook-during-the-covid-19-pandemic-a-changed-world

https://www.who.int/publications/m/item/draft-landscape-of-covid-19-candidate-vaccines

https://doi.org/10.1101/2020.09.03.280446

20

7. Folegatti, P.M., et al., Safety and immunogenicity of the ChAdOx1 nCoV-19 vaccine against 618 SARS-CoV-2: a preliminary report of a phase 1/2, single-blind, randomised controlled trial. 619 Lancet, 2020. 620

8. Jackson, L.A., et al., An mRNA Vaccine against SARS-CoV-2 - Preliminary Report. N Engl J 621 Med, 2020. 622

9. Zhu, F.C., et al., Safety, tolerability, and immunogenicity of a recombinant adenovirus type-5 623 vectored COVID-19 vaccine: a dose-escalation, open-label, non-randomised, first-in-human 624 trial. Lancet, 2020. 395(10240): p. 1845-1854. 625

10. Querec, T.D. and B. Pulendran, Understanding the role of innate immunity in the mechanism 626 of action of the live attenuated Yellow Fever Vaccine 17D. Adv Exp Med Biol, 2007. 590: p. 627 43-53. 628

11. Querec, T.D., et al., Systems biology approach predicts immunogenicity of the yellow fever 629 vaccine in humans. Nat Immunol, 2009. 10(1): p. 116-125. 630

12. Kasturi, S.P., et al., Programming the magnitude and persistence of antibody responses with 631 innate immunity. Nature, 2011. 470(7335): p. 543-7. 632

13. Thanh Le, T., et al., The COVID-19 vaccine development landscape. Nat Rev Drug Discov, 633 2020. 19(5): p. 305-306. 634

14. Jackson, N.A.C., et al., The promise of mRNA vaccines: a biotech and industrial perspective. 635 NPJ Vaccines, 2020. 5: p. 11. 636

15. Hassett, K.J., et al., Optimization of Lipid Nanoparticles for Intramuscular Administration of 637 mRNA Vaccines. Mol Ther Nucleic Acids, 2019. 15: p. 1-11. 638

16. Zeng, C., et al., Formulation and Delivery Technologies for mRNA Vaccines. Curr Top 639 Microbiol Immunol, 2020. 640

17. Fuller, D.H. and P. Berglund, Amplifying RNA Vaccine Development. N Engl J Med, 2020. 641 382(25): p. 2469-2471. 642

18. Chan, K.R., et al., Cross-reactive antibodies enhance live attenuated virus infection for 643 increased immunogenicity. Nat Microbiol, 2016. 1: p. 16164. 644

19. Chan, C.Y., et al., Early molecular correlates of adverse events following yellow fever 645 vaccination. JCI Insight, 2017. 2(19). 646

20. Chan, K.R., et al., Metabolic perturbations and cellular stress underpin susceptibility to 647 symptomatic live-attenuated yellow fever infection. Nat Med, 2019. 25(8): p. 1218-1224. 648

21. Salti, S.M., et al., Granzyme B regulates antiviral CD8+ T cell responses. J Immunol, 2011. 649 187(12): p. 6301-9. 650

22. Ehrchen, J.M., et al., The endogenous Toll-like receptor 4 agonist S100A8/S100A9 651 (calprotectin) as innate amplifier of infection, autoimmunity, and cancer. J Leukoc Biol, 652 2009. 86(3): p. 557-66. 653

23. Yu, G., et al., APRIL and TALL-I and receptors BCMA and TACI: system for regulating 654 humoral immunity. Nat Immunol, 2000. 1(3): p. 252-6. 655

24. Groom, J.R. and A.D. Luster, CXCR3 in T cell function. Exp Cell Res, 2011. 317(5): p. 620-656 31. 657

25. Conticello, S.G., et al., Interaction between antibody-diversification enzyme AID and 658 spliceosome-associated factor CTNNBL1. Mol Cell, 2008. 31(4): p. 474-84. 659

26. Premkumar, L., et al., The receptor binding domain of the viral spike protein is an 660 immunodominant and highly specific target of antibodies in SARS-CoV-2 patients. Sci 661 Immunol, 2020. 5(48). 662

27. Rogers, T.F., et al., Isolation of potent SARS-CoV-2 neutralizing antibodies and protection 663 from disease in a small animal model. Science, 2020. 664

28. de Alwis, R., et al., Impact of immune enhancement on Covid-19 polyclonal hyperimmune 665 globulin therapy and vaccine development. EBioMedicine, 2020. 55: p. 102768. 666

29. Honda-Okubo, Y., et al., Severe acute respiratory syndrome-associated coronavirus vaccines 667 formulated with delta inulin adjuvants provide enhanced protection while ameliorating lung 668 eosinophilic immunopathology. J Virol, 2015. 89(6): p. 2995-3007. 669

30. Hashem, A.M., et al., A Highly Immunogenic, Protective, and Safe Adenovirus-Based 670 Vaccine Expressing Middle East Respiratory Syndrome Coronavirus S1-CD40L Fusion 671


https://doi.org/10.1101/2020.09.03.280446

21

Protein in a Transgenic Human Dipeptidyl Peptidase 4 Mouse Model. J Infect Dis, 2019. 672 220(10): p. 1558-1567. 673

31. Higgins, B.W., L.J. McHeyzer-Williams, and M.G. McHeyzer-Williams, Programming 674 Isotype-Specific Plasma Cell Function. Trends Immunol, 2019. 40(4): p. 345-357. 675

32. Wec, A.Z., et al., Longitudinal dynamics of the human B cell response to the yellow fever 17D 676 vaccine. Proc Natl Acad Sci U S A, 2020. 117(12): p. 6675-6685. 677

33. Tam, H.H., et al., Sustained antigen availability during germinal center initiation enhances 678 antibody responses to vaccination. Proc Natl Acad Sci U S A, 2016. 113(43): p. E6639-679 E6648. 680

34. Cirelli, K.M., et al., Slow Delivery Immunization Enhances HIV Neutralizing Antibody and 681 Germinal Center Responses via Modulation of Immunodominance. Cell, 2019. 177(5): p. 682 1153-1171 e28. 683

35. von Herrath, M.G. and A. Bot, Immune responsiveness, tolerance and dsRNA: implications 684 for traditional paradigms. Trends Immunol, 2003. 24(6): p. 289-93. 685

36. Jin, B., et al., Immunomodulatory effects of dsRNA and its potential as vaccine adjuvant. J 686 Biomed Biotechnol, 2010. 2010: p. 690438. 687

37. Pepini, T., et al., Induction of an IFN-Mediated Antiviral Response by a Self-Amplifying RNA 688 Vaccine: Implications for Vaccine Design. J Immunol, 2017. 198(10): p. 4012-4024. 689

38. Grifoni, A., et al., Targets of T Cell Responses to SARS-CoV-2 Coronavirus in Humans with 690 COVID-19 Disease and Unexposed Individuals. Cell, 2020. 181(7): p. 1489-1501.e15. 691

39. Mateus, J., et al., Selective and cross-reactive SARS-CoV-2 T cell epitopes in unexposed 692 humans. Science, 2020. 693

40. Le Bert, N., et al., SARS-CoV-2-specific T cell immunity in cases of COVID-19 and SARS, 694 and uninfected controls. Nature, 2020. 695

41. Maruggi, G., et al., Engineered alphavirus replicon vaccines based on known attenuated viral 696 mutants show limited effects on immunogenicity. Virology, 2013. 447(1-2): p. 254-64. 697

42. Thompson, J.M., et al., Mucosal and systemic adjuvant activity of alphavirus replicon 698 particles. Proc Natl Acad Sci U S A, 2006. 103(10): p. 3722-7. 699

43. Ramaswamy, S., et al., Systemic delivery of factor IX messenger RNA for protein replacement 700 therapy. Proc Natl Acad Sci U S A, 2017. 114(10): p. E1941-E1950. 701

44. Gan, E.S., et al., Dengue virus induces PCSK9 expression to alter antiviral responses and 702 disease outcomes. J Clin Invest, 2020. 703

45. Ayouba, A., et al., Multiplex detection and dynamics of IgG antibodies to SARS-CoV2 and the 704 highly pathogenic human coronaviruses SARS-CoV and MERS-CoV. J Clin Virol, 2020. 129: 705 p. 104521. 706

46. Atyeo, C., et al., Distinct early serological signatures track with SARS-CoV-2 survival. 707 Immunity, 2020. In Press. 708

47. Tan, C.W., et al., A SARS-CoV-2 surrogate virus neutralization test based on antibody-709 mediated blockage of ACE2-spike protein-protein interaction. Nat Biotechnol, 2020. 710

48. GeurtsvanKessel, C.H., et al., An evaluation of COVID-19 serological assays informs future 711 diagnostics and exposure assessment. Nat Commun, 2020. 11(1): p. 3436. 712

713 714


https://doi.org/10.1101/2020.09.03.280446

22

FIGURES 715

716 Figure 1. Design and Expression of a SARS-COV-2 vaccine with conventional mRNA 717 and self-transcribing and replicating RNA (STARR®) platforms. A) Schematic diagram 718 of the SARS-CoV-2 self-replicating STARR RNA (LUNAR®-COV19) and conventional 719 mRNA vaccine constructs. The STARR construct encodes for the four non-structural 720 proteins, ns1-ns4, from Venezuelan equine encephalitis virus (VEEV) and the unmodified 721 full-length pre-fusion spike (S) protein of SARS-CoV-2. The mRNA construct also codes for 722 the same SARS-CoV-2 full length spike S protein. B) Physical characteristics and RNA 723 trapping efficiency of the LNP encapsulating conventional mRNA and LUNAR-COV19 724 vaccines. C) Western blot detection of SARS-CoV-2 S protein following transfection of 725 HEK293 cells with LUNAR-COV19 and conventional mRNA. D) In vivo comparison of 726 protein expression following IM administration of LNP containing luciferase-expressing 727 STARR RNA or conventional mRNA. Balb/c mice (n=3/group) were injected IM with 0.2 728 µg, 2.0 µg and 10.0 µg of STARR RNA or conventional mRNA formulated with the same 729 lipid nanoparticle. Luciferase expression was measured by in vivo bioluminescence on days 730 1, 3 and 7 post-IM administration. S domain 1 = S1, S domain 2 = S2, transmembrane 731 domain = TM, cytoplasmic domain = CP; aka = also known as. 732


https://doi.org/10.1101/2020.09.03.280446

23

733 Figure 2: Clinical Scores, mouse weights and transcriptomic analysis of immune genes 734 following vaccination with LUNAR-COV19 or conventional mRNA SARS-CoV-2 735 vaccine candidates. A) C57BL/6 mice (n=5/group) were immunized with either PBS, 736 mRNA or LUNAR-COV19 (doses 0.2 µg, 2 µg or 10 µg), weight and clinical scores 737 assessed every day, bled at day 1 post-immunization, sacrificed at 7 days post-vaccination 738 and lymph nodes harvested. Gene expression of inflammatory genes and immune genes were 739 measured in whole blood (at day 1) and lymph nodes (at day 7), respectively. B) Expression 740 of IFN and inflammatory response genes in whole blood presented as heatmap of z scores. C) 741 Lymph node weights at 7 days post-vaccination. Principal component analysis (PCA) of 742 immune gene expression following vaccination with conventional mRNA or LUNAR-743 COV19 at doses D) 0.2 µg, E) 2 µg and F) 10 µg. Volcano plots of fold change of LUNAR-744 COV19 versus conventional mRNA (x-axis) and Log10 P-value of LUNAR-COV19 versus 745 conventional mRNA (y-axis) for doses G) 0.2 µg, H) 2 µg and I) 10 µg. Study design 746


https://doi.org/10.1101/2020.09.03.280446

24

schematic diagram created with BioRender.com. Weights of lymph nodes were compared 747 between groups using a two-tailed Mann-Whitney U test with * denoting 0.05<P<0.01. 748

749 750 Figure 3. Cellular immune responses following vaccination with LUNAR-COV19 and 751 conventional mRNA. C57BL/6 mice (n=5 per group) were immunized with 0.2 µg, 2.0 µg, 752 or 10.0 µg of LUNAR-COV19 or conventional mRNA via IM, sacrificed at day 7 post-753 vaccination and spleens analyzed for cellular T cell responses by flow-cytometry and 754 ELISPOT. A-B) CD8+ and C) CD4+ T effector cells were assessed in vaccinated animals 755 using surface staining for T cell markers and flow-cytometry. D-E) IFNγ+ CD8+ T cells and 756 F) Ratio of IFNγ+/ IL4+ CD4+ T cells in spleens of immunized mice were assessed following 757 ex vivo stimulation with PMA/IO and intracellular staining. G-I) SARS-CoV-2 S protein-758 specific responses to pooled S protein peptides were assessed using IFNγ ELISPOT assays 759 following vaccination with mRNA (H) or LUNAR-COV19 (I). Percentage of CD8+ cells, 760 CD4+ cells, IFNγ and IL4 producing T cells were compared between groups using two-tailed 761 Mann-Whitney U test with * denoting 0.05<P<0.01, and **0.01<P<0.001. 762 763 764 765 766 767


https://doi.org/10.1101/2020.09.03.280446

25

768 769 Figure 4: LUNAR-COV19 elicits a higher quality humoral response than conventional 770 mRNA platform. A) BALB/c and C57BL/6J mice were immunized via IM with 0.2 µg, 2 771 µg, or 10 µg of LUNAR-COV19 or conventional mRNA (n=5/group). Blood sampling was 772 conducted at baseline, and days 10, 19, 30, 40, 50 and 60 post-vaccination for BALB/c and 773 days 10, 20 and 30 for C57BL/6J. B-C) IgM and D-E) IgG against the SARS-CoV-2 S 774 protein over time, assessed using insect cell-derived whole S protein in a Luminex immuno-775 assay (measured as MFI). IgG endpoint titers to mammalian-derived whole S protein, S1, S2 776 and RBD proteins to mammalian-derived whole S protein at day 30 post-vaccination were 777 assessed in F) BALB/c and G) C57BL/6J. H) Avidity of SARS-CoV-2 S protein-specific 778


https://doi.org/10.1101/2020.09.03.280446

26

IgG at day 30 post-immunization was measured using 8M urea washes. I) Neutralizing 779 antibody (PRNT50 titers) at day 30 post-vaccination against a clinically isolated live SARS-780 CoV-2 virus measured in both BALB/c and C57BL/6J. Gray dashed lines depict the serum 781 dilution range (i.e. from 1:20 to 1:320) tested by PRNT. J) PRNT50 and K) PRNT70 of 782 SARS-CoV-2 neutralization at day 30 and day 60 post-vaccination in BALB/c and 783 convalescent sera from COVID-19 patients. L) Correlation analysis of Spike-specific IgG 784 endpoint titers against SARS-CoV-2 neutralization (PRNT50). Antibody data were compared 785 between groups using a two-tailed Mann-Whitney U test with * denoting 0.05<P<0.01, and 786 **0.01<P<0.001. 787 788 789 790


https://doi.org/10.1101/2020.09.03.280446

27

791 792 Figure 5. LUNAR-COV19 elicits Th1 biased immune responses. SARS-CoV-2 spike-793 specific IgG subclasses and the ratio of IgG2a/c/IgG1 at 30 days post-vaccination with 794 LUNAR-COV19 and conventional mRNA in A) BALB/c and B) C57BL/6J mice. Th2 795 cytokine and Th1/Th2 skew in CD4 T cells at day 7 post-vaccination in C57BL/6J mice 796 measured by ICS as C) percentage of IL4+ CD4 T cells and D) ratio of IFNγ+/IL4+ CD4+ T 797 cells. Antibody titers and T cell data were compared between groups using a two-tailed 798 Mann-Whitney U test with * denoting 0.05<P<0.01, and **0.01<P<0.001. 799 800 801 802 803 804 805


https://doi.org/10.1101/2020.09.03.280446

28

806 807 Figure 6. Single dose of LUNAR-COV19 protects hACE2 mice against a lethal challenge 808 of SARS-CoV-2 virus. A) hACE2 transgenic mice were immunized with a single dose of 809 either PBS or 2 µg or 10 µg of LUNAR-COV19 (n=5 per group), then challenged with live 810 SARS-CoV-2 at 30 days post-vaccination, and assessed for either survival (with daily 811 weights and clinical scores) or sacrificed at day 5 post-challenge and measured lung and 812 brain tissue viral loads. Study design schematic diagram was created with BioRender.com B) 813 Live SARS-CoV-2 neutralizing antibody titers (PRNT70) measured at 28 days post-814 vaccination. C) Weight, D) clinical score and E) survival was estimated following challenge 815 with a lethal dose (5x10^5 TCID50) of live SARS-CoV-2 virus. F) Viral RNA and G) 816 infectious virus in the lungs and brain of challenged mice were measured with qRT-PCR or 817 plaque assay, respectively. PRNT70 and viral titers (RNA and plaque titers) were compared 818 across groups using the non-parametric Mann-Whitney U test. Weights and clinical scores at 819 different timepoints were compared between PBS and 10ug LUNAR-CoV19 immunized 820 mice using multiple t-tests. P-values are denoted by * for 0.05<P<0.01, ** for 0.01<P<0.001, 821 *** for 0.001<P<0.0001, **** P<0.00001. 822 823 824


https://doi.org/10.1101/2020.09.03.280446

29

825 826

827 828 Supplementary Figure 1. Whole blood transcriptomic data at 1-day post-prime vaccination 829 showing Nanostring counts per 50ng RNA of selected IFN and inflammatory genes. 830 831 832 833


https://doi.org/10.1101/2020.09.03.280446

30

834

835 Supplementary Figure 2. Correlation analysis of live SARS-CoV-2 neutralization against 836 binding IgG and IgG subclasses in BALB/c and C57BL/6J mouse strains. A) Spearman 837 correlation analysis of SARS-CoV-2 neutralization (PRNT50) against total IgG specific to 838 several SARS-CoV2 antigens, including S, S1, and RBD recombinant proteins. B) Spearman 839 correlation analysis of SARS-CoV-2 neutralization (PRNT50) against SARS-CoV2 S-840 specific IgG subclasses (IgG1 and IgG2a or IgG2c). 841 842


https://doi.org/10.1101/2020.09.03.280446

Documents

Title: 2 Vaccine Produces Protective Adaptive Immunity In ... · 03/09/2020 · Immune gene expression following LUNAR-COV19 and conventional mRNA vaccination . 123 . C57BL/6J mice